Time Series with PyTorch

4

PyTorch Fundamentals

Having examined how to evaluate forecasting models, we now move from assessment to construction. The previous chapter established how we quantify predictive reliability; this one establishes the elements of neural networks required to build models capable of achieving it.

Such models require understanding the mathematical principles that guide the construction of neural network algorithms and the tools that bring those principles to life. For this reason, our next step is to learn the basic building blocks of PyTorch that are fundamental to the deep-learning approaches explored throughout the book.

PyTorch provides a transparent environment for experimentation, allowing us to construct, test, and adapt neural networks ...